It really depends on what you're interested in solving for and how you want to use the results. There are three ways you could do this analysis:
- Meshing each gear in its entirety with 3D solid elements.
- Meshing each gear in its entirety with 2D plane stress elements.
- Cutting out a portion of each gear and meshing it with either 2D plane stress or 3D elements.
If the thickness of the gears is relatively small compared to the other dimensions (think of a cog on a bicycle), then a 2D Plane Stress assumption is good. If the gear is relatively thick compared to the other dimensions, then 3D will be more appropriate. How thick are the gears?
Cutting out a portion of each gear is, technically speaking, not correct; the constraints that you'd add on the cut surfaces to prevent theta rotation will result in a model that is stiffer. This increased stiffness is probably not too big of an issue (if your cut planes are far enough away from the gear teeth you're interested in), but the only real way to quantify the difference would be to solve both. Cutting in the red lined area would be (in my opinion) way too close; I'd suggest cutting out a wedge of each gear that is about 90 degs.
I got satisfactory results using the third option.
I only want to add one information I needed some time to discover. Maybe very obvious for some people but not for me or other with the same knowledge of SW like me.
The cutting option is called split.
agree with shaun here, i don't see any way to leverage symmetry or cyclic symmetry or submodeling in anyway. best to go with 2d. but if you have elements that aren't being stressed, why not just make the mesh over there really coarse?
This could be another good option, but I had some problems to define which part I want for a coarse meshing and which one a fine meshing.